Graduate School of Informatics

The study aimed to classify task difficulties using wavelet transform images of EEG signals and deep learning models. The EfficientNet-B0 model achieved the highest accuracy, but its performance varied significantly across individuals and task difficulties, indicating limited generalizability. The study suggests a need for further research to improve model generalizability, optimize performance, and validate the models on larger, more diverse datasets.

Date: 06.09.2023 Place: MODSIMMER Meeting Room

New technologies in distributed energy systems offer solutions for managing demand and generation variability in the grid. Trade between small grids allows cost reduction and system flexibility. This thesis proposes Multi-Agent Reinforcement Learning model for short-term energy trading among peers. It integrates short-term forecasts for load, generation, and price, leading to more accurate decision-making. Real-world data simulations demonstrate improved efficiency and stability compared to rule-based agents. This approach empowers prosumers to respond effectively to dynamic energy markets, enhancing grid reliability, energy efficiency, and sustainability.

Date: 07.09.2023 / 11:30 Place: -

In this study, changes in neural activation during joint action were investigated with hyperscanning, using functional near infrared spectroscopy (fNIRS) and electroencephalography (EEG) as participants performed the same task individually and then collaborated as a dyad. 62 participants were tested in dyads with a dual n-back task. Results are in support of the social facilitation model, and neural measures including hyperscanning were successful to differentiate between individual and social task settings. Overall, this study is one of the first studies to conduct EEG and fNIRS-hyperscanning during a complex collaboration task that brought together temporal constraints and elements of visuospatial reasoning.

Date: 05.09.2023 / 11:00 Place: A-212

Despite the high mortality and morbidity rates associated with alcohol use disorder (AUD) worldwide, the incomplete understanding of its complex molecular mechanism is the biggest limitation in providing successful treatments to patients. Monozygotic twin studies have found that AUD is about 50-60% heritable, and they have also revealed that environmental factors and epigenetic mechanisms play a role in the etiology of this disorder. In this study, we investigated excessive alcohol consumption related DNA methylation changes in four brain regions and blood in wistar rats to contribute to understanding to the understanding of the etiology of AUD.

Date: 04.09.2023 / 09:30 Place: B-116

This study aims to provide a bibliometric map of Functional Near Infrared Spectroscopy (fNIRS), a growing research area worldwide, in terms of interdisciplinary and institutional collaboration. Interdisciplinary and institutional collaborations are increasingly encouraged by universities as they have a high academic citation impact. In this study, the number of articles, number of citations, CNCI and IREW percentage in the field of fNIRS from 1980 to 2020 were analysed. Interdisciplinary and university & country collaborations that can increase these academic indicators are analysed from a bibliometric perspective.

Date: 11.09.2023 / 10:30 Place: -

This PhD thesis delves into the impact of selective directional auditory attention on eye movements and pupil responses. Concentrating on single-ear stimuli is shown to enhance attentiveness and task performance. The research explores effects of multi-talker speech interference, age, gender, and audio direction on eye movements and pupil dilation. Notably, interference influences pupils, prompting gaze shifts toward sound sources and larger pupils during competing speakers. This underscores the study's significance. These findings highlight the value of eye movement patterns and pupillometry in understanding auditory attention nuances. The implications are profound, benefiting hearing aid technology, communication for autism spectrum disorder and hearing-impaired individuals. The study also uncovers links between speech quality, word familiarity, and emotional connotations on pupil responses, contributing to a comprehensive comprehension of auditory processes.

Date: 06.09.2023 / 10:00 Place: A-212

This study explores the transformative potential of virtual reality in laparoscopic surgical training. Traditional box trainer implemented as a VR simulation, eliminating the need for physical tools and improved with incorporating feedback from surgeons and participant questionnaires. Results revealed that the VR version, equipped with tutorials, haptic feedback, and assisted grabbing physics, was more usable and accepted. Kinematic analysis showed similarities to physical training, and physiological monitoring revealed increased heart rate and reduced heart rate variability during tasks. This multimodal approach highlights VR's potential to enhance laparoscopic surgical skill development, providing an immersive and realistic training experience.

Date: 01.09.2023 / 10:00 Place: A-212

This study aims for detection and DOA estimation of gunshot sound onboard a drone, despite the excessive ego-noise. Array Correlation Map concept is introduced for improved detection through unanimity among sensors of an array. Also, adaptive auto-tuning to advantegous CWT scales brings adaptive denoising for transient events of varying frequency. Although studied specifically for the processing of gunshot sounds on drones, the novelties this study offers is expected to generalize to other array processing applications. Results reveal signal-to-noise-ratio enhancement, successful muzzle and shock wave signal detection, and DOA estimation performance improvement.

Date: 04.09.2023 / 10:00 Place: A-212

As number of proteins increases, computational methods play vital role for accurate function annotation. In spite of these methods, there are issues to compete which are reliable datasets and fair evaluation of performances. To address them, a fair comparison tool, ProFAB, Open Protein Functional Annotation Benchmark is developed. It aids computational and experimental researchers with its easy access to datasets and machine learning algorithms for protein function prediction using Gene Ontology terms and Enzyme Commission numbers.

Date: 21.08.2023 / 11:00 Place: A-212

A new method has been developed to study allosteric protein regulation, which is important for understanding how proteins function. The method represents proteins as networks and identifies allosteric pathways, or sites where molecules bind and regulate protein activity, in order to predict the presence and location of allosteric regions. This could potentially aid in the development of targeted therapies.

Date: 28.07.2023 / 10:00 Place: A-212